In a milestone achievement for industrial artificial intelligence, Aramco and Yokogawa Electric Corporation successfully deployed multiple autonomous control AI agents at the Fadhili Gas Plant in Saudi Arabia (on October 29, 2025). This revolutionary reinforcement learning-based technology led to significant improvements in operational efficiency and reduced energy consumption.
- Introduction: Artificial Intelligence at the heart of the energy industry
- What is autonomous control artificial intelligence?
- The challenge: Optimizing acid gas removal (AGR) processes
- Technology used: The innovative FKDPP algorithm
- Implementation phases and safe training
- Impressive results: Efficiency and sustainability
- The future of autonomous operations in industry
- FAQs
Introduction: Artificial Intelligence at the heart of the energy industry
Yokogawa Electric Corporation announced a historic milestone in deploying industrial AI solutions in collaboration with Aramco, a world-leading integrated energy and chemicals enterprise. This achievement represents a qualitative leap in how complex industrial facilities are managed and operated, successfully implementing autonomous control AI agents at the Fadhili Gas Plant in Saudi Arabia.
What is autonomous control artificial intelligence?
Autonomous control AI is defined as an artificial intelligence system capable of independently deriving optimal control methods while possessing sufficient flexibility to adapt to new and unexpected situations. Rather than following pre-programmed instructions or relying on static control models, this type of AI learns from operational data and makes real-time decisions to optimize complex operations, thereby reducing the need for continuous human intervention.
The challenge: Optimizing acid gas removal (AGR) processes
The solution implemented at the Fadhili plant focuses on autonomous control and the optimization of acid gas removal (AGR) operations. This process is essential in natural gas processing to remove harmful acidic components such as hydrogen sulfide and carbon dioxide to make the gas safe for use and transport. However, optimizing this process presents a significant challenge due to its complexity and high consumption of energy and chemicals.
Technology used: The innovative FKDPP algorithm
The AI solution utilizes multiple coordinated AI agents based on an advanced reinforcement learning algorithm known as «Factorized Kernel Dynamic Policy Programming» (FKDPP). Reinforcement learning is a type of machine learning where an agent learns how to act in an environment through trial and error to maximize rewards.
The FKDPP algorithm was developed in collaboration with the Nara Institute of Science and Technology in Japan and received Japan’s prestigious Prime Minister’s Award in 2023 for its innovation and ability to handle complex dynamic systems.
Implementation phases and safe training
The project was executed with great care across three phases, gradually optimizing process sections until autonomous control of the core AGR process was fully achieved. To ensure safety and reliability—which is critical in the oil and gas industry—a high-fidelity digital twin simulator of the plant was built to train the AI agents in a virtual environment. The system’s reliability was rigorously evaluated before integration into Yokogawa’s CENTUM VP operational control system.
Impressive results: Efficiency and sustainability
Preliminary results show significant benefits exceeding initial expectations. The system achieved a notable 10% to 15% reduction in the usage of amine (the chemical used for absorption) and steam (used for regeneration). It also achieved a 5% reduction in overall energy consumption.
In addition to cost savings and a reduced environmental footprint, autonomous control dramatically improved process stability and reduced the need for human operator intervention, even under changing ambient conditions such as temperature fluctuations.
The future of autonomous operations in industry
Khaled bin Yazid Al-Qahtani, Aramco’s Senior Vice President of Engineering Services, stated that the company is deploying industrial AI to enhance efficiency, sustainability, and value, and that this collaboration reinforces its technological leadership.
Meanwhile, Kunimasa Shigeno, President and CEO of Yokogawa, expressed pride in implementing this technology, noting that the results align with Yokogawa’s transition toward autonomous operations. This achievement represents a major step toward a future where complex industrial facilities operate independently, continuously adapting to achieve maximum efficiency, safety, and sustainability.
FAQs
Q: What is industrial artificial intelligence?
A: It is the application of AI technologies, such as machine learning and reinforcement learning, to optimize processes in industrial environments such as manufacturing, energy production, and chemical processing.
Q: What is reinforcement learning?
A: It is a type of AI where the system learns to make decisions through trial and error, receiving “rewards” for good decisions and “penalties” for poor decisions, with the goal of improving performance over time.
Q: How does this system enhance sustainability at the gas plant?
A: By optimizing processes to reduce the use of chemicals (such as amine) and energy (steam and electricity), the system reduces the facility’s environmental footprint and greenhouse gas emissions associated with energy production.
Q: Does this system completely replace human operators?
A: The system aims to reduce the need for continuous manual intervention in routine and complex tasks, allowing human operators to focus on higher-level tasks and strategic oversight of overall operational safety.
